agh, my bad. had the wrong environment up so was running 1.7.0. 1.8.0 works for me

FierceHamster54 did you find a way around this? i’m also looking for a solution here without skipping the decorators and building the Task/PipelineController manually

example pipeline
` from typing import List

from clearml import PipelineDecorator

@PipelineDecorator.component(
execution_queue="default",
docker="python:3.7",
packages=["clearml"]
)
def download(files: List[str]):
if isinstance(files, str):
print("its a str")

for f in files:
    print(f"download file: {f}")

@PipelineDecorator.pipeline(
name='something',
project='batch data processing',
version='0.1'
)
def batch_pipeline():
return download(
...

1.8.0